Striking Disparities in ER Evaluation of Chest-Pain

MILWAUKEE, Feb. 2 -- When patients with chest pain arrive at the emergency department, they may get profiled, before decisions are made on care, by race, female gender, and insurance coverage, found a national study by researchers here.

A retrospective study of 7,068 patients, corresponding to more than 32 million such visits annually, found differences in who gets ECGs and x-rays as well as cardiac rhythm and oxygen saturation monitoring using pulse oximetry, said a study published in the February issue of Academic Emergency Medicine.

This nationwide study may provide a way to explore links between existing in-patient and population-based studies on disparities in cardiovascular care, reported Lilliana Pezzin, Ph.D., J.D., of the Medical College of Wisconsin here, and colleagues at Johns Hopkins

These data, she and colleagues said, augment inpatient and population-based studies on race, gender, and insurance disparities in evaluation for coronary artery disease. Inpatient research, they said, is limited in that its starting point is generally after a diagnosis has been made. At the same time, population-based studies leave unanswered questions concerning the entrance of patients into the health care system.

The data in the ER research were drawn from the U.S. National Hospital Ambulatory Health Care Survey of Emergency Departments (NHAMCS-ED). Visits were made from 1995 to 2000 for persons 30 years or older presenting with chest pain.

The rate of ER visits for patients with chest pain increased in the six year period paralleled by an increase in the disparities in care, the researchers reported.

Overall, African American men were 25% to 30% less likely to receive ECGs and x-rays as well as cardiac rhythm and oxygen saturation monitoring using pulse oximetry than non-African American men (white or other race).

Also being a woman, whether African American or non-African American, jeopardized care.

The researchers noted that because administrative data often underestimate visits by persons of different ethnic groups, such as those of Hispanic descent, they were unable to examine trends in test ordering among other population subgroups. Thus the white and Hispanic groups were labeled non-African American.

Insurance coverage affected treatment, as well, the researchers reported. Patients who were uninsured, self-paying, or who had noncommercial insurance, such as Medicaid, or who had missing insurance data also received less comprehensive care, the researchers reported.

The adjusted probability of ordering a test was highest for non-African American patients for all four tests considered.

African American men had the lowest probabilities (74.3% and 62% for electrocardiography and chest radiography, respectively), compared with 81.1% and 70.3%, respectively, for non-African American men. These differences were significant at the P<0.02 level.

Only 37.5% of African American women received cardiac monitoring, compared with 54.5% of non-African American men.

Similarly, African American women were significantly less likely than non-African American men to have their oxygen saturation measured.

Female gender was a handicap regardless of race, the researchers found. Both African American women and non-African American women had lower rates for most tests during the study period, with an increase in these disparities from 1995 to 2000.

Patients who were uninsured or used self-pay, as well as patients with so-called "other" insurance, also had a lower probability than insured persons of having these tests ordered. For example, compared with patients covered by commercial insurance, persons covered by non-commercial insurance (Title V, other government insurance) were 13.4% less likely to have an ECG, 20.8% less likely to be placed on cardiac monitoring, 23% less likely to have oxygen saturation measured, and 13.4% less likely to receive chest radiography.

The researchers also found that about 82% of commercially insured non-African American men received an ECG, a 26.7% higher proportion than that for uninsured African American men and a 31% higher proportion than African American men covered by non-commercial insurance. These differences represent a 24% and 31% increase, respectively, over disparities prevailing in 1995.

The results for other tests reveal a similar pattern of widening disparities between 1995 and 2000 across race, gender, and insurance groups, the researchers said.

These results provide compelling evidence of a persistent, and sometimes widening, pattern of gender, racial, and insurance disparities in the ordering of noninvasive tests commonly used in the early evaluation of patients with chest pain.

Discussing the study's limitations, the researchers noted that although information on racial background, as broadly defined for this study, is relatively reliable within discharge and abstract data, reliance on these definitions of ethnic origin tends to be problematic.

Furthermore, they said, evaluation of emergency department patients with chest pain is a very complex decision-making problem. This study evaluated only four of these diagnostic variables and controlled for only a limited number of potential confounders and thus certainly did not provide a complete picture of how these decisions were made. In addition, information on laboratory tests, which would be of great interest was not available.

Finally, they said, although this study contributes to the growing literature on disparities in the provision of cardiac care, examining the data did not permit them to determine the underlying cause or causes, the appropriateness, or the clinical impact of the observed differences in test utilization. As a consequence, they wrote, these results do not lead to any direct conclusions concerning either the cause or the appropriateness of the disparities observed.

Despite these qualifications, the investigators said that these observed differences should catalyze further study into the underlying causes of disparities in cardiac care at an earlier point of patient contact with the health care system.

Identifying the root causes and their relative contributions to disparities in emergency department care is arguably both a more critical and more difficult task than documenting the existence of these disparities, Dr. Pezzin and colleagues said. For example, some patients may actually be at lower risk of having ischemic disease and therefore be treated differently. On the other hand, she said, no such explanation can account for insurance-status disparities.

Isolating the exact circumstances and degree to which inappropriate direct health-care provider bias contributes to observable disparities in the emergency department or health care in general remains a necessary but elusive goal, the investigators wrote.

"Our findings, however, underscore the need for further research to go beyond the inpatient setting when attempting to understand the causes of racial differences in the care of patients with ischemic heart disease," they concluded.